Temperature fluctuations, physical stresses, and other environmental conditions affect the fibers of an optical communication system. In particular, these factors give rise to fiber birefringence, which can unpredictably change the states of polarization (SOP) of an optical signal traveling in a fiber, especially in a single-mode fiber. These changes in the SOP are manifested as fading of the optical signal at the output end of a fiber (i.e., polarization dependent loss (PDL)) and, in some instances, polarization mode dispersion (PMD) loss.
To correct the polarization state of optical signals emerging from an optical fiber, some conventional polarization controllers typically transform the output polarization states of optical signals into prescribed or preferred polarization states for a specific application, such as interferometric signal processing or PMD compensating. Using well-known algorithms, other conventional polarization controllers can transform any arbitrarily varying input SOP of an optical signal into any arbitrary output SOP by either rotating wave-plates or varying the phase retardation of wave-plates.
To effectuate endless tracking and control of SOP for an optical signal, one approach requires an electro-optical polarization controller to include a reset cycle when the controller's operating range is exceeded. But these reset cycles generally gives rise to periods of unacceptable data loss. Therefore, some SOP controllers operate in a limited control range with occasional resetting to obtain a complete range of SOP control for optical signal transmission, which minimizes the loss of a polarization state of a local optical signal or information.
Some other types of conventional polarization controllers provide endless and continuous control of SOP almost over an infinite range rather than being restricted to a limited range of operation. These types of controllers have been designed to include cascaded polarization transformers, each having a limited transformation range. Examples of polarization transformers are fiber squeezers and electro-optic devices using lithium niobate or liquid crystal wave-plates. While these combined, cascaded devices permit substantially endless (reset-free) operation overall, individual constituent elements within traditional polarization control devices still require occasional reset cycles.
Although the reset cycles can be performed without affecting the overall polarization transformation (i.e., quasi-endless polarization control), these devices generally require complicated, computer-controlled driving algorithms for proper operation, which generally results in a slow response to fluctuations in SOP. A common approach to keep output SOP invariant uses additional variable wave-plates with liquid crystal or fiber squeezer-based polarization controllers with a computer or processor to control the phase retardation of the wave-plates. The computer calculates how best to satisfy a pair of equations for resetting the driver voltage to an initial value, and then limits electrical-driving devices to a specific range of operation. In this approach, a reset operation occurs only when the driver voltage reaches a maximum limit. This approach, however, tends to result in slow resetting of polarization controller elements, which in turn results in suboptimal control of SOP.
A first type of traditional polarization controller uses an electrical-field to control liquid crystal (LC) cells as variable wave-plates. FIG. 1A illustrates an example of this type of polarization controller. Polarization controller 100 includes four cells 101, 102, 103, and 104, where slow (i.e., horizontal) axes 105 and 107 of respective cells 101 and 103 are either parallel or perpendicular to each other, and the slow axes of cells 106 and 108 are oriented at ±45 degrees to the axis of cell 101. A computer-controlled driving algorithm or a switchable, double optical path is used for resetting controller 100, which occurs at a relatively slow resetting speed. A drawback to this approach is that four electrical LC drivers (shown as drivers 1, 2, 3 and 4) are required for this type of polarization controller to transform any arbitrarily varying input into any arbitrary output.
Liquid crystal-based polarization controller devices are widely used as phase modulation devices. Liquid crystals include fluids that derive their anisotropic physical properties from the long-range orientational order of their constituent molecules. Also, liquid crystals exhibit birefringence, which is a function of the orientation of the liquid crystal molecules. The orientation can be controlled by the intensity of an applied electric field. For a normal liquid crystal used as a phase retarder, the phase retardance, δ, depends on the liquid crystal layer thickness, d, and birefringence, Δn, as expressed in Equation 1. That is:
                              δ          =                                    2              ⁢                                                          ⁢              π              ⁢                                                          ⁢              d              ⁢                                                          ⁢              Δ              ⁢                                                          ⁢              n                        λ                          ,                            (                  Equation          ⁢                                          ⁢          1                )            where λ is the wavelength of the incident light. For a half-wave plate, δ=π.
Reorientation of the liquid crystal molecules under the influence of an applied field introduces elastic strains in the material. These strains stem from constraints imposed on the molecular orientation at the boundaries confining the liquid crystal. These surface constraints are given the term “surface anchoring.” In most practical applications, the surface anchoring is such that molecules close to a surface are not free to reorient, but rather remain substantially along some preferred direction.
When an electric field is applied to a liquid crystal element, such as, a homogenously-aligned half-wave plate, the directors of LC molecules are reoriented in response to the applied field. Typically, the response time is usually ˜1 ms, depending on the properties of the LC. During the response time, the phase retardance of the half-wave plate is a non-linear function of time. Generally, substantially similar cells show similar time functions.
A second type of traditional polarization controller uses fiber squeezers to mechanically induce birefringence in the fiber axes, which in turn causes retardation between the two orthogonal modes perpendicular and parallel to the direction of pressure. FIG. 1B shows an example of this type polarization controller. Polarization controller 150 includes first to fifth fiber squeezers 160, 162, 164, 166, and 168, each comprising a pair of piezo-electric actuators 154 and 156. Polarization controller 150 further comprises a single mode optical fiber 152 that receives side pressures from each pair of piezo-electric actuators 154 and 156 to generate birefringences. Polarization controller 150 also includes a control unit 170. Control unit 170 includes an A/D converter 176, a microprocessor unit 174 and a D/A converter 172 for driving fiber squeezers to change the SOP of an optical signal. A drawback to polarization controller 150 is that it uses five drivers, each of which requires resetting. Another example of this type of polarization controller uses a rotatable fiber clamp to supply the necessary retardation and optical axis orientation. But because this controller requires mechanical movement for its control, polarization fluctuation in transmission fibers typically cannot be controlled in real time.
A third type of polarization controller provides substantially reset-free, endless polarization transformations from any arbitrarily varying optical input polarization into any arbitrarily output polarization. This type operates by producing adjustable elliptical birefringence of constant total phase retardation in a single-mode fiber. A particular transformation is obtained by adjusting the azimuth of linear birefringence and the ratio of linear-to-circular birefringence. Structurally, this type of controller is made up of three controllable fractional wave elements (i.e., plates) in cascaded combination. To realize endless polarization transformations, the orientations of optical-axes of the fractional wave plate elements are controlled such that the fractional wave elements function the same as three cascaded rotating wave-plates (such as an endlessly rotatable half-wave element and two synchronously rotatable quarter-wave elements). This type of polarization controller can be realized using either distributed bulk optic devices or integrated electro-optic waveguide devices. Proper rotation of the wave elements is afforded by using a feedback control circuit to monitor the outputted optical polarization, and then to generate an appropriate electrical drive signal to achieve the proper rotation. Although this type of polarization controller operates sufficiently for most of its intended functions, it does not provide suitable wavelength and temperature independence.
Besides the relatively slow resetting of conventional liquid crystal-based polarization controllers, the drawbacks associated with the above-mentioned polarization controllers include, among other things, relatively high cost, elevated operating voltages, mechanical fatigue, and high insertion loss.
In view of the foregoing, a polarization controller having a fast resetting capability is highly desirable. Ideally, the polarization controller would be an inexpensive, highly responsive device for controlling SOP of optical signals emerging from optical fiber systems, and would have improved polarization mode dispersion compensation for high-speed, optical communication systems.